Project Summary/Abstract More than a third of human cerebral cortex responds to visual information, which enters cortex through primary visual cortex (V1). Accordingly, much is known about V1 processing. However, much less is known about information processing in the `next' cortical area (V2), and even less is known about higher area V3. In animal studies of V2, one key to understanding information processing is that different types of stimulus features are processed ~independently in segregated `stripes' of cortical columns. One set of columnar stripes responds selectively to variations in color, a surface property. Another set of columns responds best to variations in boundaries based on stereoscopic (`3D') stimuli (thus the latter research can clarify mechanisms underlying amblyopia). Study of such columns can reveal fundamental responses in the component neurons. Previously, it has been unknown whether such functional stripes/columns exist in humans. Conventional functional magnetic resonance imaging (fMRI) techniques cannot spatially resolve such columns, which are quite small. By using a specialized fMRI approach (7T, 1 mm3), here we successfully demonstrated (and have begun analyzing) such columns and stripes in human areas V2 and V3. Beyond simply imaging such columns, we propose to test a new hypothesis: early visual cortex processes boundaries vs. surface features, in two parallel streams, within thin and thick stripes (respectively) in V2. We also test whether that surface-vs.- boundary information is then passed forward to segregated columns in V3. If confirmed, this hypothesis will clarify (and help unify) the diverse existing data about the nature of functional processing in early visual cortex. Aim 1 will test our hypothesis by presenting different types of stimulus boundaries or surface properties, testing for correspondingly higher responses in thick or thin stripes, respectively. Each of five sub-aims will isolate and test responses to boundaries defined by differences in either: binocular disparity (sub-aim 1.1), or direction of motion (1.2), texture (1.3), color (1.4), or luminance (1.5). Aim 2 tests the conclusions from Aim 1 in a complementary way, by presenting stimuli that include only a single boundary, in each quadrant of the visual field, and in the corresponding cortical map. Aims 3.1 and 3.2 will test for higher activity in retinotopically-predicted subregions in thick stripes (and related V3 columns) due to inferred or illusory contours. The stimulus used in aim 3.3 predicts a shift in activity between the thin vs. thick stripe systems, produced by retinotopically specific colored surfaces- vs. boundaries, respectively. Different points in visual space (either surfaces or boundaries) are mapped onto the cortical surface, a property known as `retinotopic' mapping. Thus the columns (aim 1) co-exist with maps of retinotopy. However It is not known how these two maps accommodate each other at the spatial resolution of columns. Aim 3 will clarify whether thin vs. thick stripes (and related V3 columns) have duplicated retinotopic maps (sub-aim 3.1), or specific retinotopic `gaps' (3.2), or smaller vs. greater retinotopic spread, respectively.